7 Requirements of a Modern BI Platform

  • Milena Marinkovic
  • February 12, 2018
7 Requirements of a Modern BI Platform

Gartner said it all when they completely re-defined the Magic Quadrant for BI and Analytics earlier this year. “The shift to the modern BI and analytics platform has now reached a tipping point,” said Ian Bertram, managing vice president at Gartner. “Organizations must transition to easy-to-use, fast and agile modern BI platforms to create business value from deeper insights into diverse data sources.”

So, what should you look for in a Modern BI and analytics platform? Independent analyst Neil Raden shares his thoughts in his new paper, Evolution of BI in the Era of Big Data:

  1. The total abstraction of the end-to-end process so analysts can do their work and not get bogged down or delayed by system complexity. The more analysts can concentrate on their own work in an environment where the tools fit together seamlessly, the more likely it is that the population of analysts will grow.
  2. Extreme performance from distributed, commodity clustered architecture. With the bulk of the processing executing on high-performance clusters, in the cloud or on-premises, the expansive data pools can be leveraged efficiently.
  3. Extreme agility. By not being limited to physical data models in relational databases or proprietary cube architectures, the scope of analytics is vastly expanded. New business questions and use cases can be addressed without re-modeling or duplicating data.
  4. Robust business modeling capability. BI tools are weak at modeling. Excel is the de facto standard for building models and performing what-if and scenario modeling. Attaching a spreadsheet interface to big analytics restores modeling (as opposed to reporting) to the analysts’ workbench and provides an easy, familiar interface to apply analytics.
  5. Statistical and quantitative analysis using vast amounts of data. A great deal of analysis in organizations is related to advanced functions such as graph and path analysis, along with data and compute-intensive algorithms that were not practical before “big data.” While traditional BI analysis was limited to filtering and simple arithmetic, evolved big data analytics offers sentiment analysis, clustering, decision trees, graph analysis, and other powerful techniques.
  6. Limitless types of data sources. Traditional BI was limited to structured, highly aggregated data. Today’s explosion of data is coming from many diverse sources and formats, requiring new analytic systems to incorporate data with no structure, ragged hierarchies, many-to-many relationships, or newer structures like graph models.
  7. RESTful API’s. This allows app libraries to be used with little to no coding. Apps can be added to enrich analysis without integration effort. Vendors and third parties can “curate” apps, providing a level of security and accurate operation.